TY  - RPRT
TI  - You Can Learn Tokenization End-to-End with Reinforcement Learning
AU  - Sam Dauncey
AU  - Roger Wattenhofer
PY  - 2026
UR  - https://arxiv.org/abs/2602.13940
ID  - 2602.13940
ER  -